Baker: UWM has been developing cutting edge tech for quite some time. You have Leo, MIA, Chat UWM — a chatbot, virtual assistant, and loan estimate optimizer. Some of the people I’ve spoken to about AI have emphasized that it’s “use it or lose it” — and by “lose it,” they mean your job. That might sound overly dramatic, but the idea is that this is something that’s here to stay. It’s becoming more and more powerful. You and your company are at the vanguard of all of this — defining how it works for everybody.

Here are a few things I’d love to know more about: What is UWM doing with artificial intelligence? How is AI shaping the industry? And what do NMP readers — originators, brokers, executives, bankers, many of whom are aware of AI but aren’t necessarily using it yet — need to know?

Bressler: Overall, AI is as revolutionary to the mortgage industry as the paperless revolution was about twelve years ago — as big as the internet was in the early ‘90s.

We’re at a point where if loan officers, companies, vendors — if they don’t adopt AI and really understand what it means for their workflow and business process — they’re going to get left behind. It’s like not creating a website in the early ‘90s and thinking everything would just happen through referrals and word of mouth. Five years later, no one can find you, and the demographics have changed. That’s exactly what’s happening in the mortgage industry right now.

From an adoption standpoint, the biggest thing people need to understand is that too many folks are trying to make a faster horse. They’re using AI just to say, “I’m doing something with AI,” and that’s the wrong way to look at it. There is no faster horse at this point.

You’ve got to build something revolutionary. Build a car. Build the Model T. AI gives you the ability to examine your entire workflow. And most of us have been in the industry at least a decade, because we were riding the wave of low interest rates. Now those rates have risen, and we have the time — not always the money, but the time — and the experience to say, “Okay, let’s rethink how AI can really change and enhance what we do.”

Because that’s what AI should be doing. It should not — and right now it does not — replace people’s jobs.

AI itself won’t do that — but people who use AI will. That’s the difference.

If you’re going to be strategic and stand at the forefront, you’ve got to pick a lane. That’s the easiest way to do it. You can look at the borrower experience — how do I get more borrowers to my site, how do I get that information into my system? You can look at CRMs — how am I using my CRM to stay in front of leads, especially for the stuff I don’t want to do? How can I use AI to sell better?

Then there’s the operations piece.

At UWM, we’ve done all of those things. A lot of our secret sauce in becoming the number one mortgage company in the country was a massive technology-focused initiative. But we did it very altruistically — and I really mean that. We built all this technology for our broker partners. Then we gave it away for free.

We told them, “You don’t even have to use us.” Obviously, we want you to — we’ve got great rates, great customer service, amazing products — but at the end of the day, we believe that choice is better. And brokers represent that choice.

A lot of those shops are small — one or two people, maybe ten. They don’t have the budget to compete with Chase or these larger retail lenders. So we said, we’re going to level the playing field for you.

If we support brokers first, and they understand what we’re doing, they’ll come back into the ecosystem. Then they really have a chance — not just to compete, but to win.

From an AI adoption standpoint, because we gave away tools like ChatUWM, Mia, and LEO, they’re already using them. They don’t have to go out and learn new platforms.

Because we build everything ourselves, and because we give it away for free, we’re basically forcing adoption back into the market. If other people copy us — great. That helps the industry. It makes all of us more efficient.

There are lots of lanes you can check into. There are tons of vendors out there who would beg for your business if you don’t have the capacity to build these tools. And they definitely understand their own workflow. They know how they’re using CRMs, how they’re attracting borrowers to their website, or their storefront, or whatever it is. And that’s the way people need to be looking at AI.

Baker: What stands out is how you’re emphasizing the win-win. These tools serve brokers and LOs, but they also serve the client. Who wouldn’t want to get a call saying they could save money?

Let me ask: Does Mia identify herself as AI? Or does she leave that ambiguous?

Bressler: She will absolutely identify herself if you ask her. We build that into the system. Either we disclose it up front or we make it available on request.

That said, she sounds very human. The conversations feel natural, and most people don’t realize they’re speaking to AI unless they ask directly.

Her job is to assist and enhance the LO’s work. As we gather more data, we’ll continue evaluating whether to lead with the fact that she’s AI. But if you ask, “Are you a robot?” or “Are you AI?” she’ll say, “I’m a virtual assistant. I work for Andy. I can help answer your questions right now.”

Funny enough, when we launched ChatUWM and Mia in pilot groups, the most common question users asked was, “Do you have a boyfriend?” Or, “Can you be my girlfriend?”

She handles it all gracefully. She’ll say, “Hey, let’s keep it professional.” You can ask her for her Instagram and she’ll respond, “I’m not comfortable with that. Can we get back on topic?”

She’ll literally talk to you about anything, but she’ll always keep it appropriate.

Baker: I imagine you have to “red team” her behavior — run tests to see how she responds to all kinds of questions. Some of it’s AI-generated, but some of it must be programmed, right?

Bressler: Exactly. That’s all prompt engineering. You test everything and create structured responses for the edge cases.

Baker: And you’re gathering a lot of data, too — seeing where the friction points are, the common questions. I imagine people’s mortgage-related struggles are fairly universal, but hearing how they articulate their concerns gives you insight into how to refine the tool.

Bressler: Totally. We’ve always relied on gut instinct in this industry. For decades, we’ve guessed what borrowers want and how to reach them. That’s worked to an extent.

Now we have real data. Conversations. Patterns. Specific questions. It’s powerful.

We can say, “Okay, these are the top three concerns. Let’s build a product that addresses them.” Or, “These are the knowledge gaps. Let’s train our LOs to fill them.”

Baker: I’m curious — what kind of feedback are you getting from people who do realize Mia is AI? Not just borrowers, but also brokers and LOs.

I think back to early ChatGPT days, when people would say, “This is clunky” or “It hallucinates.” There was resistance. But as it improved, those barriers started to fall.

At the same time, I read a study that said: as AI tools got stronger, customers were more likely to say, “I still want a human in the loop.”

So as AI capabilities have grown, so has the desire for a human voice. How has that insight shaped your journey? And how have brokers reacted?

Bressler: If I look at ChatUWM specifically, one of the reasons it works so well is because of our customer service. That’s been a core strength of UWM for 30+ years.

At the end of the day, no matter what anyone says — whether it’s millennials, Gen Z, or Gen X — everybody wants to talk to a human. Especially in mortgages. It’s the biggest financial decision most people will ever make.

So I tell people all the time: AI is not going to take your job. But if you never adopt any AI tools or learn how they work, someone else will — and they will take your job.

UWM built these tools to put critical information at people’s fingertips instantly. You could call in to UWM, talk to an underwriter, run a scenario — that might take 15 minutes of back and forth.

Now, you can go into ChatUWM, maybe while riding in the passenger seat, and get the same answer in under a minute. Then you can call the client back. It’s not replacing the underwriter. It’s just giving you faster access to the information you need.

That’s how we approached this: not to replace human interaction, but to make people more efficient.

Even with Mia, you can transfer to a real person at any time. That’s the point. She’s there to give you work-life balance, to answer common questions, to make those calls you should be making but don’t have time for.

And one thing that sets us apart at UWM is: we don’t do retail. We’re not trying to take business from brokers. We’re not trying to collect borrower data and run our own funnel.

Our tech exists to return business to wholesale loan officers. That’s our approach. We don’t want to take things away — we want to give LOs more control.

We’re not trying to change human behavior. People will be doing mortgages long after I retire. We’re just focused on making the process more efficient.

Baker: As these tools continue to become more powerful — which is clearly where it’s all heading — you’ve got ChatUWM answering questions, LEO feeding LOs information in real time, and Mia making simultaneous phone calls.

As that power grows, where do you see it all going next? I’m sure there are projects you can’t talk about yet, but directionally, how is this going to shape the industry?

Bressler: I think the industry’s going to evolve in a few phases.

Right now, the first wave is already here: larger vendors and lenders — those that made early investments in AI — are using it to accelerate access to data and decisioning.

If you look at product and pricing engines, a lot of them have done an excellent job narrowing down the options using AI.

From a mortgage tech standpoint, it’s really just us and maybe one or two others at that level. That’s Phase 1.

The second wave will be widespread adoption. Whether people realize it or not, they’re already using AI. Unless you’re still handwriting your 1003s, you’re interacting with it in some way.

And what excites me most is the third wave: personalization.

As more people understand how AI works, it’s going to become incredibly easy to create custom AI agents. We’re almost there already. People just don’t realize how accessible it is.

Soon, LOs will be able to build AI agents tailored to their exact workflow, the way they like to do business. That kind of individual control is going to explode.

It’s great for the industry — but even better for borrowers. Because it gets them the best deal, faster. And when borrowers win, the whole American economy wins.

Baker: What’s interesting about what you’re describing is that in this next phase — the “agentic age” — not only will brokers, LOs, and companies have AI agents, but borrowers will too.

Everyone along the pipeline will have an assistant helping them through the process. So you’ll start seeing agents talking to agents. I’m fascinated to see how that plays out.

I want to be mindful of your time, so I’ll end with one final question. For the people who read this story or watch the clips — mostly brokers, LOs, folks across the industry — if you had a moment to whisper something in their ear, what’s the one thing they really need to take away from this conversation? What do they need to understand about AI in the mortgage industry?

Bressler: They need to understand that the industry has already moved into AI. This isn’t some future-state thing — it’s here, right now.

And no matter how you start — start. Use ChatGPT if you have to. Use ChatUWM. A lot of the people listening probably already have chatbots or AI tools they use individually.

The key is: you’ve got to start somewhere. You’ve got to understand the capabilities of AI, and just as importantly, you’ve got to understand how you do business.

I gave a speech a few weeks ago and changed my entire message after listening to some of the other presenters. What I realized is this: everyone is coming for all of the business.

All the big players are targeting all of the pie.

So if you’re an LO who’s been in the game a while, and you’ve built a good book of business — don’t just give it away for some short-term gain. Because if you do, you will lose it.

You’ve worked too hard. You’ve nurtured these relationships. You’ve stayed up late. You’ve missed time with your family. You’ve earned what you have.

AI is here to enhance every part of your job. And only you know how you work best. Only you can pause, take a step back, and ask, “How can I do this better?”

There are a million AI products out there that can improve your life. But you need to start right now.

Baker: Fantastic. Thank you. These are important insights. And I 100% agree. It might sound trite, but it really is now or never. This is happening. It’s not about making peace with AI — it’s about embracing it and seeing how it can help you.

Bressler: That’s a great way to put it. Absolutely.